Bug 753703

Summary: Xorg gets into busy loop in SIGIO handler
Product: [Fedora] Fedora Reporter: Zdenek Kabelac <zkabelac>
Component: xorg-x11-serverAssignee: X/OpenGL Maintenance List <xgl-maint>
Status: CLOSED RAWHIDE QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: rawhideCC: xgl-maint
Target Milestone: ---   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2012-04-20 20:01:33 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
busy looping xserver none

Description Zdenek Kabelac 2011-11-14 08:35:13 UTC 
Created attachment 533467 [details]
busy looping xserver

Description of problem:

After recent fix of my reported bug 737031 I'm now experiencing busy looping Xserver. Unfortunately I do not have any exact way how to trigger this, but it already happened to me twice. I think the fix for bug 737031 introduced some regression.

I'm adding backtrace from the moment the Xorg busyloops - I should note, that in this moment the server cannot be killed in other way then '-9' (thus in fact I'm back at the same moment as prior fix for bug 737031 as console is again unusable)

Version-Release number of selected component (if applicable):
xorg-x11-server-Xorg-1.11.99.1-1.20111109.fc17.x86_64

How reproducible:
Not yet know an exact steps.

Steps to Reproduce:
1.
2.
3.
  
Actual results:


Expected results:


Additional info:
Comment 1 Zdenek Kabelac 2011-11-20 11:25:18 UTC 
It might be worth to attach also this trace from Xorg.0.log:

---

[mi] EQ overflowing.  Additional events will be discarded until existing events are processed.

Backtrace:
0: /usr/bin/X (xorg_backtrace+0x2f) [0x4635ff]
1: /usr/bin/X (mieqEnqueue+0x243) [0x555d13]
2: /usr/bin/X (0x400000+0x47183) [0x447183]
3: /usr/bin/X (xf86PostButtonEvent+0xf7) [0x48fd47]
4: /usr/lib64/xorg/modules/input/synaptics_drv.so (0x7fb82ae34000+0x221d) [0x7fb82ae3621d]
5: /usr/lib64/xorg/modules/input/synaptics_drv.so (0x7fb82ae34000+0x33cd) [0x7fb82ae373cd]
6: /usr/lib64/xorg/modules/input/synaptics_drv.so (0x7fb82ae34000+0x4df3) [0x7fb82ae38df3]
7: /usr/bin/X (0x400000+0x7f6c8) [0x47f6c8]
8: /usr/bin/X (0x400000+0xa4b6b) [0x4a4b6b]
9: /lib64/libpthread.so.0 (0x7fb82e1e0000+0xf500) [0x7fb82e1ef500]
10: /lib64/libc.so.6 (writev+0x1b) [0x7fb82cf3115b]
11: /usr/bin/X (0x400000+0x6ef5c) [0x46ef5c]
12: /usr/bin/X (0x400000+0x666f4) [0x4666f4]
13: /usr/bin/X (FlushAllOutput+0x17c) [0x466f7c]
14: /usr/bin/X (0x400000+0x33b1a) [0x433b1a]
15: /usr/bin/X (0x400000+0x22fc5) [0x422fc5]
16: /lib64/libc.so.6 (__libc_start_main+0xed) [0x7fb82ce6a73d]
17: /usr/bin/X (0x400000+0x232b1) [0x4232b1]
[mi] These backtraces from mieqEnqueue may point to a culprit higher up the stack.
[mi] mieq is *NOT* the cause.  It is a victim.
[mi] EQ overflow continuing.  100 events have been dropped.
...
Comment 2 Zdenek Kabelac 2011-11-20 11:46:49 UTC 
It looks like the trace in comment 1 is just something which happened already many times in past - i.e. bug Bug 465884.
Comment 3 Zdenek Kabelac 2011-11-21 14:34:22 UTC 
I've also create separate thread on lkml with  kernel back trace which seems to be related to this issue as well.


https://lkml.org/lkml/2011/11/21/151
Comment 4 Adam Jackson 2012-02-09 18:12:03 UTC 
Blocked on writev?  How odd.  What is that file descriptor pointing to?
Comment 5 Zdenek Kabelac 2012-03-13 10:52:00 UTC 
Since I'm now using kernel 3.2/3.3 - I've not experienced such problem for quite some time - so it might have been already fixed - I guess if there will not be a similar problem within another month it might be closed as fixed with newer a kernel.
Comment 6 Adam Jackson 2012-04-20 20:01:33 UTC 
Closing per comment #5, not seen any other reports like this.